The Jaipur Garden Carpet
The Safavid ‘garden’ carpet in Jaipur’s Albert Hall Museum was shown in its entirety for the first time in HALI 171, Spring 2012. Scroll down to see it in its full glory, following the HALI article by Bruce Healy:
‘The formal four-part garden as a traditional representation of Paradise probably originated in ancient Mesopotamia during the Sumerian period. It evolved through the walled hunting parks of the Achaemenians – the garden built by Cyrus the Great at the 6th century BC site of Pasargadae, near Shiraz in southwestern Iran, is the oldest known – and those of their Sasanian successors. It later became intertwined with the descriptions of Paradise in the Qur’an as Islam spread across the Middle East and Central Asia.
Perhaps the earliest literary references to a ‘garden’ carpet are to the so-called ‘Spring Carpet’ that graced the enormous audience hall of the 7th century Sasanian emperor Chosroes (Khusrow) II in his palace at Ctesiphon. Following the Arab conquest in 637 AD, this carpet was said to have been cut into pieces and distributed as booty to his troops at the behest of the Caliph Omar. After that, as far as we know, the ‘garden’ carpet theme lay dormant for about a thousand years.
Antique Persian carpets in the chahar bagh (four-garden) design have long held a special status among collectors. Apart from a tiny handful of technically distinct late 16th/ early 17th century carpets attributable to Safavid south Persia (Kerman), all known examples, woven in wool on a cotton and/or wool foundation, were made during the 18th century in workshops in northwest Persian Kurdistan. This production died out by the early 19th century.
By the early years of the 20th century collecting interest was focused on the small number of surviving 18th century Kurdish pieces. James Ballard had one, Hagop Kevorkian and Joseph McMullan two each. Lord Aberconway had the greatest of all (plus a substantial fragment). Collectors as prominent as Otto Bernheimer were reduced to stitching together inconsequential fragments to make a mat in something resembling chahar bagh format.
Then, in 1937, a long forgotten chahar bagh carpet was discovered in a locked storage room in the abandoned palace of the Maharajas of Jaipur at Amber Fort. Knotted in the manner associated with carpet production in Safavid Kerman, this great early carpet had inventory records on a label attached to the lining, as well as notes written directly on the lining itself. The earliest inscription provided a brief description of the carpet along with the date 29 August 1632. Thus we know that it was made no later than the first third of the 17th century, perhaps earlier.
The format of the Jaipur carpet is a bird’s-eye view of the rigid rectilinear plan of the most strictly construed fourgarden layout. The field is divided into quarters by a broad longitudinal watercourse and a somewhat narrower transverse watercourse. These primary watercourses intersect at a large, nearly square pool, and in the centre of the pool is a pavilion shown in elevation. The walls of the pavilion are alternately red brick and tile, with a crenellated parapet and a blue-tiled onion dome. The interior walls are decorated with tile mosaics and murals depicting naturalistic flowers and leaves on stems, with the addition of two fantastic horned quadrupeds in the squinches. In the centre of the pavilion is an elaborate throne.
The top and bottom halves of the field are each bisected by a narrower secondary transverse watercourse, each intersecting the longitudinal watercourse in a rectangular pool much smaller in size than the central pool.
All the primary and secondary watercourses and pools are bordered by frieze-like avenues planted with an assortment of trees (among them planes, date palms, and fruit trees in blossom or bearing peaches, oranges, or pomegranates), shrubs, and flowers (including roses, carnations, and lilies). Against the backdrop of the foliage are foraging pheasants and waterfowl; fluttering, flirting, and nesting songbirds (some feeding their young); gazelles, some startled and others stately; and the occasional feline. Mammals and birds are depicted in profile, trees, shrubs, and flowers in elevation. This same formula was used in virtually all known chahar bagh carpets.
The remaining portions of the field, separated from the primary and secondary watercourses by the avenues, are divided into square garden plots, arranged in groups of four (at the ends of the field) or five (near the central pool) and irrigated by a tertiary system of water channels that, in turn, intersect in hierarchically smaller pools and connect the entire water system with a stepped meander around the periphery.
The garden plots have their own assortment of trees, shrubs and flowering plants, birds, and an expanded menagerie of mammals. Canines and felines chase gazelles and each other. A bear cub confronts a large bird. Some less recognisable land mammals bear a rather unlikely resemblance to kangaroos. Each plot represents an independent scene, with the trees (sometimes only one, more often three to five) oriented vertically, horizontally, and even diagonally. The one constant is that, as in the avenues lining the watercourses, the trees always spring from the bank of a watercourse or pool. A consequence of this approach is that the orientation is constantly changing so that there is no single vantage point from which to get a consistent view of more than a small fraction of the design.
Water in the pools and channels is indicated by short, wavy, blue-green lines against an ivory ground that is presumed to represent white marble beneath the water. This representation of the water appears in only one other garden carpet, the small Figdor rug at the Museum für angewandte Kunst in Vienna. In all others, including the only other great early century Kerman garden carpet known, the Wagner carpet in the Burrell Collection, Glasgow, the water is indicated by a stylised representation of the coloured tiles lining the bottoms of the pools and channels.
The watercourses and pools in the Jaipur carpet are stocked with a single species of fish (perhaps trout), swimming single file or sometimes paired in the primary watercourses and pools. In the larger watercourses and pools they are joined by waterfowl (occasionally biting fish or each other), and in the pools on the primary longitudinal watercourse by pairs of mythical qilin in combat, as well as eels and turtles, and, in the central pool, a single dragon biting and being bitten by fish, plus one lone, unidentifiable, terrestrial quadruped that seems to have been lured into the central pool by the opportunity to nibble on fish. A grouping that looks like several snails or chi surrounding a cloud-band hovers over the dome of the pavilion. Throughout the watercourses and irrigation channels, the fish are joined, and occasionally replaced, by aquatic plants. Just as on land, all creatures in the water are shown in profile.
Compared with the early Safavid ‘Paradise Park’ and ‘Hunting’ carpets, which are usually assigned to the 16th century, the terrestrial ambiance here is idyllic. There are no animal combat scenes, and even the interactions between apparently wild animals seem almost playful. Violence is confined to the water.
There is a lot going on in this carpet. The avenues and garden plots are made up of vignettes that have a strong resemblance to the repeated scenes in many contemporaneous carpets with landscape designs. As Michael Franses has pointed out, the strong similarity of composition among the scenic designs used during the Shah Abbas period in carpets from several different weaving areas, as well as in textiles, miniature paintings, and lacquered book bindings, suggests that they were derived from cartoons drawn by a small number of artists. Kerman carpets with landscape designs normally have either centralised or directional designs, and consequently provide a satisfying viewpoint from either the centre or one end of the carpet. In contrast, the constant change of orientation across watercourses and around pools in the Jaipur carpet is dizzying.
One simplification that helps make sense of the design is the near absolute mirror image symmetry about the primary watercourses, both transversely and longitudinally. This degree of symmetry is almost unheard of in real Persian gardens, where asymmetry is an inevitable consequence of topography. In any case the four-garden theme imposes no requirement that the gardens be identical. Each of the four walled gardens of paradise described in the Qur’an has its own symbolic fruit (pomegranates, dates, figs, and olives). So it would be rather odd if the four gardens were visually indistinguishable.
Furthermore, in the most sophisticated vase-design carpets – such as the fragmented carpet now divided among museums in Berlin, Cairo, London, and Washington – that do possess a bilateral design symmetry, the colours of the mirror image design elements are varied for subtlety, and no two design elements are identical. As a result, a totally organised and controlled design (and one that represents a garden in a form quite different from the formal chahar bagh convention) conveys the seeming lack of organisation of a natural garden.
In contrast, the mirror images in the Jaipur carpet are more likely to share identical colours that do not break symmetry. Not surprisingly there are a few details, especially in the garden plots and watercourses, that defy the symmetric scheme. Most notably, the two secondary pools on the longitudinal watercourse are quite different in terms of size, shape, and creatures cavorting within. A few creatures are mirrored transversely but transformed to altogether different beasts longitudinally. Bear cubs, for example, exchange with felines in garden plots at opposite ends of the carpet.
To place the Jaipur carpet more securely in context requires a brief digression on the features that characterise other south Persian ‘vase-technique’ carpets of commensurate quality that were produced around the same time. The finest early examples, in particular those believed to have designs originating in (and possibly also implemented by) court workshops, share a very sophisticated aesthetic approach to both overall layout and detail. The layout is invariably highly structured, the drawing is refined and consistently curvilinear, and a rhythmic fluidity permeates the entire design.
In the case of the carpets in vase design, the rosettes, palmettes, and vases are organised on multiple overlapping curvilinear ogival lattices. The consistent use of overlapping planes produces great depth. Yet because the slender stems that make up the lattices are partially obscured by the floral motifs as well as leaves and scrolling vines, the rigorous system that controls the design becomes so understated that it can be difficult to follow. In addition, while the lattices and the motifs they carry have a very high degree of formal symmetry, mirror image forms appear in different colours, introducing a tension between the viewer’s expectation of exact correspondence and the reality of subtle variation. The elaborate floral motifs that dominate the design have tremendous vitality thanks to the ingenious ways the fluid curvilinear drawing is used to convey swirling motion, directional energy, or static equilibrium.
The great medallion design carpets with landscape backgrounds, such as the Stieglitz ‘Paradise Park’ carpet in the Hermitage, are no less structured and elegantly graceful. The landscape is organised not only by its own internal logic and the need to provide a consistent frame of reference for the viewer, but also by the demands imposed by the relationship of the medallion to the landscape. The multiple layering of elements – blossoming trees overlap cypresses, with the great cats roaming in between; branches overlap within each individual blossoming tree; and the medallion with its pendants floats over the centre – adds depth, just as the multi-plane lattice did in the vase design carpets.
Viewed in this context, the Jaipur garden carpet is a little jarring and flat. The rectilinear layout of the watercourses – the skeleton on which the entire design is organised as well as the defining feature that characterises all formal Persian gardens – seems out of place amid the curvilinear drawing that informs every other aspect of the carpet. The incongruities of direction, which make viewing the design so unsettling and provoked Charles Grant Ellis to call the design confused, are an inevitable consequence of the chahar bagh layout and the need to orient the vegetation with respect to the source of irrigation. The drawing is very refined, but not sublime. Its most significant weaknesses are the flatness produced by insufficient use of overlapping layers and the slightly cluttered feeling in the foliage in the avenues and garden plots as a result of a scale too small for such a large and complex carpet. For comparison, in the Wagner garden carpet, which is quite a bit smaller, the avenues of trees are close to a metre high, compared with perhaps forty centimetres in the Jaipur carpet. So the grandeur of the avenues in the Wagner carpet is missing in the more compact scale of drawing in the Jaipur carpet.
While the field design of the Jaipur carpet may be flawed, the border is superlative. Typical of early Kerman carpets, there is a narrow main border with a slender solid ivory outer guard stripe. The pencil thin reciprocal red and blue inner guard stripe is reminiscent of the similarly skinny chevron inner guard stripe on the Jekyll sickle leaf fragment. In the main border, apricot rosettes alternate with vine-leaf palmettes on a meandering vine. The vine-leaf palmettes in turn alternate between inward and outward facing and between blue and white. Centred on the vine between each rosette and vine-leaf palmette are diagonally oriented lotus palmettes, alternating direction and colour in pairs, as well as a single five-petal flower near the diagonal intersection points with each rosette and vine-leaf palmette. From the opposing diagonals of the rosettes and vine-leaf palmettes spring more substantial boughs bearing leaves and five-petal flowers (pink with white centres on those springing from the palmettes and white with gold centres on those springing from the rosettes) that echo the boughs on the blossoming fruit trees in the field.
Only occasional breaks in the otherwise inexorable rhythm and the imperfect execution of the well-conceived corner resolutions offer proof of a less than courtly origin. It is worth noting that the boughs with blossoms in this border may be the most direct predecessor known for the rectangular grid arrangements of five-petalled blossoms in the borders of many slightly later vase carpets. Conversely, the similarities of the border in the Jaipur carpet with those on some of the earliest Kerman pieces such as the Corcoran throne carpet, the Jekyll sickle leaf fragments, and the Berlin/Cairo/London/ Washington vase carpet fragments, support the dating of all these carpets to no later than the beginning of the 17th century.
What is truly surprising is that the Jaipur carpet is both exceptionally large (eight and a half metres long and almost four metres wide), and extremely finely knotted, for a vase technique carpet (over 4,700 knots per square decimetre according to May Beattie’s structural 1972 analysis, as reported by Ellis), finer than any of the vase-technique carpets she included in the structural analysis insert of her catalogue for the 1976 Sheffield exhibition. The warps are the standard Z4S ivory cotton, with alternate warps depressed. The taut wefts are Z2S fine wool in various colours. Some of the sinuous wefts are Z2S cotton or wool, while others are blue or pink silk, which Beattie pointed out is found in the best quality carpets. It also has the splendid palette typical of the best Kerman carpets. The critique above, although niggling for such a great carpet, clearly indicates that it was not the product of a court atelier. Rather, the carpet was produced by a commercial workshop, albeit one of the highest calibre and most likely as a special order.
Were there ever chahar bagh garden carpets designed entirely by court artists, never mind executed in court workshops? We will never know for sure unless one turns up some day, but our inclination is to doubt the possibility. The demands of the rectilinear chahar bagh layout seem irreconcilable with the design imperative of the Safavid court artists from whose cartoons the greatest vase carpets originated. Paradoxically, the rectilinear drawing and stylised, abstracted, design aesthetic of the 18th century northwest Persian garden carpets – usually regarded as degenerate descendants of the incomparable Jaipur carpet and its now lost siblings – provided a more congenial home for the chahar bagh format than did the more naturalistic and refined curvilinear style of the earlier Safavid period.